similar to: Out-of-sample predictions with boosting model

I'm running R 2.10.1 with mboost 2.0 in order to build predictive models . I am performing prediction on a binomial outcome, using a linear function (glmboost). However, I am running into some confusion regarding centering. (I am not aware of an mboost-specific mailing list, so if the main R list is not the right place for this topic, please let me know.) The boost_control() function allows

Hi all, First of all, I'm a novice R user (less that a week), so perhaps my code isn't very efficient. Using the MBoost package I created a model using the following command and saved it to a file for later use: model <- gamboost(fpfm,data=SampleClusterData,baselearner="bbs") # Creating a model save(model,file="model.RData") # Saving a model After this, during a

I just upgraded to Ubuntu 8.10 (i386) from 8.04. After the upgrade, I ran update.packages(.libPaths()[1]) in R to get the packages installed from source up to date too. Unfortunately, two packages could not be updated: mclust and mboost. In both cases, the error I got mentioned lf77blas. Here's the output for mboost: * Installing *source* package 'mboost' ... ** libs gcc -std=gnu99

I'm getting started in R, and I'm trying to use one of the gradient boosting packages, mboost. I'm already installed the package with install.packages("mboost") and loaded it with library(mboost). My problem is that when I attempt to call glmboost, I get a message that " Error in glmboost() : no applicable method for "glmboost" ". Does anybody have

Hi, I use a mboost model to predict my dependent variable on new data. I get the following warning message: In bs(mf[[i]], knots = args$knots[[i]]$knots, degree = args$degree, : some 'x' values beyond boundary knots may cause ill-conditioned bases The new predicted values are partly negative although the variable in the training data ranges from 3 to 8 on a numeric scale. In order to

Dear all, I am puzzled by R's memory usage when calling the blackboost function from package mboost to estimate a Gradient boosting model on a simulated dataset with 20 correlated variables and 100,000 obs. The blackboost object created by the function is only 15.3Mb, but R's memory usage increases by about 3.9Gb during the estimation of the model and the memory is not released even after

mboost contains a blackboost method to build tree-based boosting models. I tried to write my own "cost-sensitive" ada family. But obviously my understanding to implement ngradient, loss, and offset functions is not right. I would greatly appreciate if anyone can help me out, or show me how to write a cost-sensitive family, thanks! Follows are some families I wrote ngradient <-

I'm attempting to fit boosted regression trees to a censored response using IPCW weighting. I've implemented this through two libraries, mboost and gbm, which I believe should yield models that would perform comparably. This, however, is not the case - mboost performs much better. This seems odd. This issue is meaningful since the output of this regression needs to be implemented in a

Dear useRs, we are happy to announce the release of mboost 2.0-0 on CRAN: http://cran.r-project.org/package=mboost This version contains major updates and changes to the implementation of the main algorithm. Some slight changes to the user-interface where necessary. Please consult the manual and the list of CHANGES below. The package 'mboost' (Model-based Boosting) implements

Dear useRs, we are happy to announce the release of mboost 2.0-0 on CRAN: http://cran.r-project.org/package=mboost This version contains major updates and changes to the implementation of the main algorithm. Some slight changes to the user-interface where necessary. Please consult the manual and the list of CHANGES below. The package 'mboost' (Model-based Boosting) implements

Sorry for the tardy reply but I just found your posting incidentally today. To make long things short: You are right about the centering. We forgot to correct the intercept if center = TRUE. We lately found the problem ourself and fixed it in the current version (mboost 2.0-3). However the problem only occurred if you extracted the coefficients. As the intercept is rarely interpretable we

Full_Name: Ivan the Terrible Version: 2.9.2 OS: Windows XP SP3 Submission from: (NULL) (89.110.13.151) When using the method blackboost_fit of the package mboost appear following error : Error in party:::get_variables(obj at responses) : trying to get slot "responses" from an object (class "boost_data") that is not an S4 object Simple test case that produce bug:

I'm interested in further understanding the differences in using many classification trees to improve classification rates. I'm also interested in finding out what I can do in R and which methods will allow prediction. Can anybody point me to a citation or discussion? Specifically, I want to classify remotely sensed imagery where training data is extracted on class membership by the user.

I am having trouble predicting new data with a model created from package mboost: > mb1<-glmboost(as.formula(formula1),data=data_train,control=boost_control(mstop=400,nu=.1)) > f.predict<-predict(mb1,newdata=data_train) Error in scale.default(X, center = cm, scale = FALSE) : length of 'center' must equal the number of columns of 'x' Ultimately I want to predict

Hi all, I am currently updating our package gamboostLSS which depends on package mboost *and* on package gamlss.dist. From mboost we use a lot of the fitting infrastructure and from gamlss.dist we obtain the relevant loss functions (aka families) used for fitting and corresponding quantile functions. Furthermore, we use the Family() function from package mboost. However, if I depend on both

Hi all, anyone has an idea how I could fix this? \donttest{ ## Set colors from colorspace package with a fallback col <- try(colorspace::rainbow_hcl(5), silent = TRUE) %||% rainbow(5) } The problem is that this makes R CMD check freak out (http://www.r-project.org/nosvn/R.check/r-devel-linux-x86_64-debian-clang/falsy-00check.html) if the colorspace package is not declared as a dependency.

Hi I am using boosting for a classification and prediction problem. For some reason it is giving me an outcome that doesn't fall between 0 and 1 for the predictions. I have tried type="response" but it made no difference. Can anyone see what I am doing wrong? Screen output shown below: > boost.model <- gbm(as.factor(train$simNuance) ~ ., # formula +

I'm unable to open an SPSS file over my network. If I copy it to my local C:/ drive I can read it. I saved the command (in a "crib sheet" text file) in order to avoid all the typing, so I'm pretty sure I've done it before. I verified that the file I'm trying to read is OK. This is what happens: > SurveyData <-

Dear List I just downloaded and installed R 2.12.0 and then installed R Commander . First it got RCmdr and Car, and then suggested for other packages for utilizing the full functionality- I clicked yes! I got 140 packages installed!!! Cran Mirror was UCLA... Here is the list. Is this intentional- I can see some packages like snow and multicore which are desirable but quite optional.(see list

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 because of strange reasons my notebook sometimes crashes short after startup (but that's not ext3's fault, maybe mem?, when i wait several minutes it works without problems) the problem is that after 3 crashes at startup, when my notebook finally worked i got the msg: Sep 23 23:29:17 blackbox kernel: EXT3-fs warning (device ide0(3,3)):